implement some fancy text classification models by using keras
Three Dataset:
- Game Title with 10 categories
- Twitter Sentiment Analysis
- JD.com product reviews with 3 categories
Model Implementation:
- NB with TF-IDF
- SVM with TF-IDF
- MLP with OneHot
- CNN(LeNet-5)
- TextCNN
- LSTM
- Bi-GRU
- RCNN
- CLSTM
- CNN and GRU
- HAN
- FastText
- Inception and TextCNNBN: idea from here
To-Do List:
Pre-trained word embedding set download link: Glove
[PS] You can not just run the code in game review part. You need to run the code step by step yourself. And you also need to change some file path when you run the code.